African American and Latino men who have sex with men (MSM) have reported some of the highest rates of HIV seroprevalence in the U.S. By contrast, Asian and Pacific Islander (API) MSM have reported much lower rates of HIV infection. More recent data, however, indicate that API and Latino MSM engage in unprotected anal intercourse (UAI) at near equal rates, but Latino men are more likely to become infected with HIV. African American men, on the other hand, report lower rates of UAI compared to API and Latino men, but are exhibiting the highest rate of HIV infection among all three groups. Our work suggests that ethnic differences in HIV prevalence may be explained to a great degree by variations in sexual partnership patterns and that sexual risk behavior among MSM of color is influenced by social discrimination (i.e., racism, poverty, and homophobia). To our knowledge, no in-depth studies have examined sexual partnership patterns among MSM of color. Even basic descriptive information about characteristics of social networks and the manner in which networks influence partnering processes is not currently available. We propose a three-phase study to describe sexual partnership patterns and explain the association between social discrimination, social networks, sexual partnerships, and HIV risk among African American, API, and Latino MSM in Los Angeles, CA. In Phase 1, we will conduct ethnographic mapping, in-depth individual interviews (N=60), and focus groups (N=96) to explore the nature of sexual partnership formation and examine the key domains hypothesized to influence social networks, sexual partnerships, and HIV risk in our working model. In Phase 2, we will utilize qualitative data to develop measures of the constructs of interest, and test these new scales to establish their reliability and validity (N=168). Based upon these new measures and existing measures in the research literature, we will develop a quantitative survey instrument. In Phase 3, we will conduct a cross-sectional survey to describe sexual partnership patterns, and examine our working model of HIV risk. A venue-based sample of 1200 men will complete audio computer-assisted self-interviews using a standardized questionnaire developed in Phase 2. The proposed study will advance theoretical understandings of HIV risk behaviors by examining potential mechanisms (i.e., social networks and sexual partnerships) through which social discrimination impacts sexual risk among MSM of color and offer valuable insights for possible interventions involving both individual and structural changes.